How I photographed the Orion Nebula from my light polluted backyard
Some of you may be wondering if this is Photoshopped.
Short answer: no. I use GIMP instead of Photoshop.
But in all seriousness, it’s very difficult to get a good astrophoto, especially from the city, without editing. Here’s a single subexposure without any editing done (aside from basic adjustments like rotating, cropping, or image conversion).
I mean, I guess it looks cool that you could see even this much from a city backyard. But we can do way better.
In this post, I’m going to go into great detail on how I managed to capture the photo at the top of the page. This is going to be long — you’ve been warned!
Part 1: Gear
With all this in mind, a basic setup for deep-sky astrophotography contains at least a DSLR attached to a telescope mounted on an equatorial mount with a sturdy tripod. The camera’s actually not as important as you would think. Even a DSLR can do a good job at astrophotography; it just needs to be able to control the shutter speed and ISO. I can attach it to my telescope with a T-ring. There are cameras designed specifically for astrophotography that have onboard cooling and all, but they’re by no means required.
The setup I used for my photo was a fairly cheap one — as astrophotography setups go. Astrophotography is one hobby that can get very expensive.
Camera: Canon EOS 70D ($300–500 for a used one)
Telescope: Astro-Tech AT60ED ($499 including accessories like the T-ring and field flattener)
Equatorial Mount: Skywatcher Star Adventurer ($399)
Tripod: Dolica Proline (~$50)
Other accessories: Neewer Remote Shutter Release Cable (~$20)
With a total of over $1000, this might seem expensive, but there are some high-end mounts alone that go for well over this price. Luckily for me, I already had the camera and the tripod beforehand.
Part 2: How Astrophotography Works
Astrophotography is a completely different beast from daylight photography. Just as in regular photography, controlling the exposure is key. There are three main ways to adjust exposure: changing the shutter speed, the aperture, or the ISO (to the best of my understanding, this controls the camera sensor’s sensitivity).
For astrophotography, shutter speeds have to be quite long; in this picture, I used 10 second exposures for the relatively bright core so it wouldn’t get blown out, and 45 second exposures for the outlying areas of the nebula. In contrast, daylight photography generally features shutter speeds of under a second.
Aperture is a bit tricky; most lenses measure aperture in terms of focal ratio, which is equal to the focal length divided by the diameter of the aperture (the hole in the lens that lets light through). For example, a 135 mm lens with an aperture size of 24.1 mm has an f ratio of f/5.6. The aperture size primarily determines the brightness of point sources of light such as stars; a bigger aperture equates to brighter stars. Meanwhile, a lower f-ratio brightens light spread out over a larger area, such as nebulae. As it happens, an important function of telescopes is to collect as much light as possible with their large apertures. The telescope I used has an aperture of 60 mm and an f-ratio of f/6.
A high ISO leads to a brighter image, so ISO values for astrophotography also have to be decently high; I used an ISO of 1000 for the 10 second exposures and 3200 for the 45 second exposures. It’s best not to go too high (like into the 10000s), as that starts introducing more noise.
Of all of these, the shutter speed is the most problematic. The Earth is rotating, so the stars also appear to move across the sky, rotating about the celestial poles. They don’t appear to move that much, but at the high focal lengths of a telescope, the apparent movement of the stars is magnified. Even after about a second of exposure, stars will appear to trail through a telescope. Star trails can look pretty artistic, but they just smear up deep sky images. This is where the equatorial mount comes in. When the mount is aligned with the celestial pole, it will rotate the telescope so that it tracks the apparent motion of the stars.
That’s not all though. Cameras always produce some amount of random noise, and taking pictures in low light conditions only makes this problem more apparent. To remedy this, astrophotographers stack many subexposures so that some of the noise averages out. This also serves to aggregate more light from the target to get more detail in the final image. The remote shutter release cable helps automate taking all those subs so I don’t shake the telescope by pushing the shutter button. For my photo, I initially took 150 subs of length 45 seconds and 30 subs of length 10 seconds. This amounts to nearly two hours worth of light gathered from the target.
Finally, make sure to shoot in RAW image format in order to retain more data. Use daylight white balance, as this closely matches the colors our eyes would see if they were more sensitive to light.
Part 3: Acquiring the Subexposures
I headed outside with my telescope to start an imaging session on the evening of February 4. The moon wasn’t going to rise until about 11 pm. Skies were clear, and an offshore wind was slowly drying out the air — perfect conditions for astrophotography.
First, I polar aligned my mount to the celestial pole. I’m in the Northern Hemisphere, so this process was fairly simple. I just used the polar scope on the back of the mount and make sure the bright star Polaris (aka the North Star) is in the right spot. Polaris isn’t exactly on the celestial north pole, and the difference becomes important when tracking a telescope. I use an app called Polar Scope Align to determine where Polaris should be in the FOV of the polar scope.
Southern Hemisphere folks have it tougher — the celestial south pole is in the constellation Octans, which is quite faint and hard to find, especially in light polluted areas.
After I polar aligned, I could slew to my target. The Orion Nebula can be found in its namesake constellation Orion. It’s an easily recognizable constellation that’s visible to the naked eye even in light polluted areas. It can be seen in the southeast during winter evenings and rises high into the southern sky. The Orion Nebula appears as the middle “star” in Orion’s sword.
Once I located Orion’s sword in the viewfinder, I took a few short test pictures to make sure that it was framed up properly. Once everything was tightened, the mount turned on, and the target centered, I started the remote shutter release and let my telescope capture some photons.
Part 4: Processing
Now for the most involved step in the process — turning those raw files into a work of art.
But first, a disclaimer: if at any point it sounds like I don’t know what I’m talking about, it’s because I don’t. I’ve changed my processing workflow a lot over the months and it’s usually a pretty messy process. So, I’ll just cover the most important things.
The first thing to do is check all the subs. The mount doesn’t track perfectly, and sometimes the stars do end up trailing. So, I needed to throw out the exposures that have star trails. I did this manually (which was super tedious), though it turns out that some processing programs, such as Pixinsight, can do this automatically.
Then, I stacked them. I use a program called SiriL to do this. I added in the raw files and debayered them to get them in color. The files will be converted to FITS files, which is a good file format for processing data.
Then, I registered the images, i.e. aligning the stars in each image. Finally, I stacked them all together. In SiriL, there are aptly named tabs for each of these tasks, and the correct options for each are shown in the screenshots below. Why these options? Honestly, I’m not completely sure how each of them work; I just do as Google guides.
I had a lot of images, so the process took around a half hour in total. Luckily, SiriL does these tasks automatically, so I just sat back and looked at some memes in the meantime.
The result of the stacking, shown below, already shows less noise and more detail than the raw subs I started out with. Note that I only stacked the 45 second exposures. I had to repeat this process for the 10 second exposures and blend both the images into the final result. Here, I cropped out the black borders that resulted from the registration.
Right now, the image is in its “linear” state, and I haven’t made any edits to it other than stacking. Since this image was made in pretty low-light conditions, the image has pretty low contrast, and much of the fainter nebulosity appears to blend in with the background right now. A look at the histogram (the distribution of how bright the pixels are), which can be brought up in the Histogram Transformation menu under Image Processing, shows a prominent spike about a quarter of the way from the left.
This shows that most of the pixels are not all that bright, which matches what’s seen in the image. However, if I do a “stretch” on the image, darkening the dark areas like the background and brightening the other areas, such as the nebulosity, I can increase the contrast and bring out some of the hidden detail.
After SiriL adjusts the necessary parameters automatically, the spikes on the histogram become much broader. The corresponding image looks like this:
There’s already much more detail. The Running Man Nebula to the left of the Orion Nebula is now clearly visible, and so is more of the Orion Nebula itself. Unfortunately, the background has now attained a reddish color due to light pollution and there are also some gradients visible in the image. Not to mention, there’s more noise visible; because I increased the contrast of the image to bring out more of the details, I also brought out the noise.
Thankfully, there are ways to deal with these problems. SiriL has a photometric color calibration tool, which takes the parameters for the image (focal length, pixel size of the camera, coordinates of the target) and fixes the colors. I also tried removing the gradient with a tool called Background Extraction.
Now, we have the following image:
The background colors definitely look more natural, but it’s debatable how well the background extraction worked. If anything, it looks more like I got different gradients. We can also see more vertical banding now.
In any case, this was the image I brought into GIMP for some more editing. Here, I blended the long exposures shown above (which feature the core of the nebula blown out) with the short ones, which had more detail in the core but lacked detail in the outer areas, using layer masks and carefully painting the area of the core so it smoothly transitions to the outer areas. I also made more adjustments to the histogram with the curves tool, darkening the background and fixing the colors some more.
However, we still haven’t fully gotten rid of the noise or the background gradients yet (though it might be hard to see in this image).
For the noise, I used a program called Topaz Denoise AI. It costs $60, but I got a free trial (which thankfully doesn’t require credit card info). The program is pretty self-explanatory; there are sliders for the level of noise reduction, sharpness, and detail. After messing around with the sliders, I got a result I was satisfied with.
Finally, to fix the background gradients, I brought the image back into GIMP. After duplicating the layer, I used the strongest median blur possible in order to make a “map” of the gradient. The nebulae are still visible in the blurred image, so I clone stamped over them. Finally, I changed the blend mode to “grain extract”, which shows a smooth bright green background and the nebulae unchanged. After darkening the colors and removing the green cast, I was finally done.
If you read all the way through, congratulations! If not, I hope the length of this post can at least help you appreciate how complex astrophotography can be.
